Addressing stress for women entrepreneurs in Ethiopia: a randomized controlled study

The primary intervention under study is a brief Acceptance and Commitment Therapy based training designed to reduce mental distress amongst high-risk populations. The study posits that the implementation of this intervention will lead to a reduction in mental distress and probable depression. This is informed by evidence from similar interventions that have found impacts on stress and distress (Tol et al. 2020), depression, and food-insecurity (Vlassopoulos et al. 2021). Moreover, the intervention would decrease the experience of Intimate Partner Violence by improving relationship quality, self-esteem, and self-regulation (Greene et al., 2019; Iverson et al.,2011; Tol et al.,2020).

The intervention involves the provision of a self-help handbook developed by the World Health Organization: ‘Doing What Matters in Times of Stress’ followed with seven 30 minute sessions over the course of ten weeks. The book has five chapters – Grounding, Unhooking, Acting on our values, Being kind and Making room. Trained lay helpers deliver the sessions based on a guiding script to reinforce the content and exercises in the handbook. Both the self-help handbook and the sessions aim to address how individuals respond to stress. They provide strategies for individuals to disengage from challenging thoughts and ground themselves in the present moment. The intervention also emphasizes the importance of understanding and acting on one's values. These approaches are aligned with the principles of Acceptance and Commitment Therapy, which guide participants in developing psychological flexibility and resilience.



The Africa Gender Innovation Lab’s Innovations in Financing Women Entrepreneurs project adapted the handbook and designed the full intervention in collaboration with local researchers. The project piloted the intervention with 100 women in 2021 with implementation support from the Ethiopian Medical Association. The program evaluated by this study comprises a scale-up of the same intervention, using the same materials and protocols with some adaptations based on the lessons from the pilot. The adaptations involved increasing the number of weekly calls from six to seven as one of the chapters was too long to cover in one session. Further, based on the feedback received in the pilot stage, entrepreneurs were offered to invite their household members to attend the self help session with them.

2023-04-03

2023-06-25

Mental Health Outcomes
1.Kessler Psychological Distress Scale (K-6) - Respondent’s score on questions measuring distress in the previous 30 days.
2. Perceived Stress - Respondent's score on frequency of stressful feelings and thoughts in the previous 30 days
3. Depression (PHQ-9) - Score on Patient Health Questionnaire and the intensity of feelings of distress

Business Performance
1. Survival of business - Current ownership of the business
2. Number of employees - Number of employees
3. Profits - Total profits of the business during the previous month (winsorised at 95%)

Intimate Partner Violence
1. Physical or Emotional violence - Dichotomous variable indicating if respondent experienced physical violence or emotional violence in the past 12 months

Secondary mental health outcomes
1. Well-being Index - Respondent’s score for the 5 questions measuring well-being in the previous two weeks (WHO).
2. WHODAS-2.0 (Disability) - Respondent’s score on questions measuring difficulty in performing day-to-day work and social participation in the previous 30 days
3. Self esteem - Respondent’s score on questions measuring self-esteem based on Rosenberg self esteem scale

Socio emotional skills
1. Grit Scale - Respondent’s score on their perceived perseverance and passion for long term goals
2. Self Control - Respondent’s score on their level of self control
3. Self-efficacy - Respondent’s score on their level of self efficacy

Intimate Partner Violence

1. Physical violence - Dichotomous variable indicating if respondent experienced physical violence in the past 12 months
2. Emotional violence - Dichotomous variable indicating if respondent experienced emotional violence in the past 12 months
3. Attitude towards IPV - Scale measuring respondent’s agreement to wife beating

Secondary Business Performance
1. Number of hours worked in business - Number of hours worked in business previous week
2. Loans in arrears - Binary variable indicating if any loan by the respondent is in arrears
3. Savings - Savings in the previous seven days
4. Satisfaction with self-employment - Score measuring (1 to 10) satisfaction with self-employment

Entrepreneurial Outcomes

1. Entrepreneurial Orientation Index - Standardized index indicating innovativeness and risk-taking orientation of the respondent
2. Entrepreneurial Behavior Index - Standardized index indicating future planning work behavior and counter productive work behavior

We screened 6000 women via phone to identify 1200 eligible women for the study. An in-person baseline survey was then conducted with these 1200 women. After completion of the baseline survey, the study participants were randomized to either the Doing What Matters in Times of Stress (DWMTS) self-help manual intervention (n=596) or waitlist control (n=595), stratifying by city and marital status. We choose city as a stratification variable to account for potential heterogeneity across cities. We also stratify by marital status, considering some of the outcome variables, such as experience of IPV, are specific to married women. They were also allowed to invite a household member to join the sessions. This was followed by 7 phone-based sessions over 7 weeks to review handbook materials for the intervention group only.

We randomised the study participants at the individual level in four phases as batches of baseline data were received from the survey firm.

We randomised the study participants at the individual level in four phases as batches of baseline data were received from the survey firm using Stata.

Individual level randomization

No

No clusters in the study

1191 individuals

Treatment group = 596 individuals, Control Group = 595 individuals

We computed the size of the minimum detectable effect (MDE) of the experimental design with the means and standard deviations of both the treatment and control groups of our baseline dataset. 1191 individuals were interviewed for the baseline survey. The randomised assignment allocated 596 women to the treatment group and 595 to the control group. We compute the minimum detectable effect for each of the outcome variables assuming 80 percent power and a significance level (𝛼) of 0.05. The MDEs for the psychological outcomes of the study vary between 5 percent and 15 percent. Similarly, the MDEs for business profits and number of employees range between 21.5 percent and 28 percent. Moreover, the MDE for the experience of any form of Intimate Partner Violence is 25.5 percent.

Ethiopia Stress Management PAP_Final.docx